In brief
Over the past year, software leaders, boards and investors have repeatedly asked us similar questions: Is software becoming easier to replicate? Will AI-native challengers disrupt incumbents? Do engineering teams need to become smaller? How much of a premium should we pay for AI maturity?
Strong opinions on these topics are easy to find. Data is harder.
So, we went looking for it, surveying 400 European software CTOs.
The answers point to something bigger than a technology cycle. AI is no longer just changing how software is built, it is changing what makes a software company valuable. Four structural shifts run through the findings, and each starts with an assumption worth testing.
The prevailing narrative is confident: AI-native entrants, free from legacy architecture and technical debt, will sweep incumbents aside.
The pressure behind that narrative is real. More than four in ten technology leaders say that, as part of renewal conversations, customers routinely probe whether part of their product could be rebuilt or replaced using AI.
Yet visibility and displacement are not the same thing.
When we looked at where these challengers gain ground, why customers choose them and where incumbents continue to hold an advantage, the picture became more nuanced.
The question is no longer simply whether AI-native competitors are entering the market. It is what they are really taking, what remains difficult to reproduce and where incumbent software companies should invest in response.
Answering that question requires looking beyond the product itself, to the way software is designed, built and embedded into customers’ processes.
AI is already being used across the software development lifecycle, from requirements and code generation to testing, debugging and documentation.
One in three CTOs surveyed believe that in three to five years, more than half of new or materially modified production code will be AI-authored or agent-generated. This is no longer only about faster coding.
How EY can help
Learn more about how EY-Parthenon Software Strategy Consulting teams apply real-world software CTO and CPO expertise to drive value.
Read moreIf AI removes friction from implementation, what then constrains the path from product strategy to reliable software in production?
A bottleneck rarely disappears altogether. It moves. Where it moves determines whether gains in coding capacity support more ambitious roadmaps and better products or merely create new demands elsewhere in the software development lifecycle.
Follow that constraint and it leads directly to people: which capabilities software organizations need, and how those capabilities are developed.
The simple version is AI writes more of the code, so software companies need fewer engineers. The changes taking place inside engineering organizations are less straightforward.
As code generation becomes more accessible, architecture, judgment, validation and governance become more important. At the same time, many organizations are raising expectations for junior talent, even as AI supports more of the routine work through which engineers have traditionally gained experience.
This raises a more consequential question than the immediate size of the engineering team.
How do engineers develop the technical judgment required for senior roles when many of the traditional steps towards acquiring it are changing?
The answer will affect hiring, mentoring, career progression and the future pipeline of architects and technology leaders. It will also influence whether AI-enabled capacity becomes a lasting organizational advantage.
AI is already contributing meaningfully to software revenue. More than one-third of respondents report that over 25% of their software or data-enabled revenue is linked to AI capabilities, AI-generated outputs, autonomous workflows or AI agents.
But creating value and capturing it are not the same thing.
AI is increasingly embedded in products and customer workflows. Companies still need to determine how to price that value, recover the costs associated with delivering it and protect margins as usage grows.
The question for management teams and investors is what evidence would make that value durable and investable. Can AI support measurable revenue or margin improvement? Can the value it creates be monetized? And does it reinforce an advantage that competitors cannot readily reproduce?
These questions connect the economics back to the product, the delivery model and the engineering organization behind it.
Together, the four shifts describe a fundamental reset in how software companies create, defend and capture value.
The full report explores where competitive pressure is translating into actual displacement, where constraints are emerging across the software development lifecycle, how engineering talent models are changing and what is required to turn AI adoption into sustainable software economics.
If you lead, invest in, or advise a software business, the shifts described here will shape your strategy for years to come.
Special thanks to Joanna Adamska and Andreas Kroon who contributed to this article.
AI is changing more than how software is built. It is reshaping where software companies find defensibility, how engineering organizations operate and how value is captured. The winners will be those that turn AI-enabled capacity into stronger products, durable customer relevance and sustainable economics.
Read more
The Confidence Trap: Industrial Guardrails for Enterprise AI
AI seems certain but hides risk. Learn how to turn output into reliable insight with deterministic validation. Discover the Foundry model.
When intelligence takes shape: the rise of physical AI
Discover how physical AI is transforming industries and jobs across Europe. Learn why acting now is crucial—and how your organization can lead the next industrial revolution.
Transforming audit and finance: The AI revolution with Datasnipper
Discover Vidya Peters, CEO of DataSnipper, insights on AI’s impact in audit and finance, customer-first innovation, and the evolving role of auditor’s landscape.
